Musculoskeletal tissue is highly sensitive to mechanical forces. These mechanical forces can influence cell signaling (Salvi & DeMali, 2018) and drive cellular responses (Dunn & Olmedo, 2016) which, in turn, controls and regulates responses and behaviors in human physiology (Wang, 2017). One type of mechanical forces are the ones experienced during exercise.

Exercise is typically classified into resistance and aerobic exercise (Wilson et al., 2012). Resistance exercise is designed to enhance muscular strength and power (Stricker et al., 2020) while on the other hand, aerobic exercise aims to improve general endurance performance (Rothschild & Bishop, 2020). Both types of exercise promote adaptations in the musculoskeletal tissues, by altering signaling pathways in charge of protein synthesis (Coffey & Hawley, 2007). A single bout of aerobic exercise is insufficient to produce significant changes to musculoskeletal tissue. Nonetheless, one exercise session promotes transient disturbances in cellular homeostasis that, when repeated over time, result in the specific exercise-induced adaptation associated with long-term training (Coffey et al., 2009; Hughes, Ellefsen, et al., 2018).

Studies have shown the effects of exercise on the morphology of the musculoskeletal system, however, to date there is no systematic review that summarizes the effects of exercise in the morphology of tissues of the musculoskeletal system. This may help guide clinical decision making when looking to elicit changes in the musculoskeletal tissues. As a primary endpoint the aim of this systematic review is to summarize the current body of knowledge regarding the morphological adaptations of the musculoskeletal system in response to chronic exercise and, as a secondary endpoint, to identify the types of exercise used to elicit said morphological adaptations.

Methods

This review was done according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines (Page et al., 2021). PROSPERO ID CRD42020188000. Studies were eligible for inclusion if they were original articles, publication date from 2000 to 2020, done in human subjects, English as publication language, had a clear exercise intervention and as an outcome presented a morphological adaptation in the tissue studied. Search was done between July 2021 and September 2021. Databases used were Pubmed, PEDro, ClinicalKey, Proquest and ResearchGate. Studies detailing acute changes were excluded. Search strategy was done using keywords and boolean operators: ((Cartilage OR Intervertebral Disc OR Bone OR Ligament OR Tendon OR Muscle) AND Exercise). All authors contributed to the initial search and screening. FL and AM performed the complete analysis per record. In the event of a disagreement a third author (LP or JCV) was called to settle a decision. The AXIS tool was used for observational studies for risk of bias assessment (Downes et al., 2016). For experimental studies, the PEDro scale was used for risk of bias assessment (Maher et al., 2003). Records were not excluded based on the assessment. Items extracted were the following: author, year of publication, study design, tissue studied, population, intervention, comparison, and outcome.

Results

A total of 2819 records were identified (314 for intervertebral disc (IVD), 666 for cartilage, 561 for bone, 366 for ligament 513 for muscle and 370 for tendon). After full-text screening 66 records were included in the final analysis (6 for IVD, 6 for cartilage, 36 for bone, 2 for ligament, 9 for tendon and 7 for muscle). PRISMA flowchart is presented in Figure 1. León et al. (2023).

Characteristics per tissue

A full list detailing each study characteristic (author, design, population, intervention, comparison and outcomes) is presented in the Appendix.

Study design

For IVD four studies were observational in design while two were experimental; for cartilage all six studies were experimental; for bone twenty-five studies were experimental and eleven were observational; for ligament both studies were observational; for tendon four were experimental and five were observational; for muscle all seven studies were experimental.

Population studied

For IVD the population studied were people that partake in sports (4) and individuals with pain (non-specific chronic low back pain (1) and radicular leg pains due to lumbar disc herniation(1)); for cartilage they were asymptomatic marathon beginners (1), women aged 45-55 years of age (1) and participants with osteoarthritis (OA)(4); for bone they were healthy participants (8), children (7), postmenopausal women (5), premenopausal women (3), participants with spinal cord injury (SCI)(2), men with prostate cancer (1), breast cancer survivors with treatment-related menopause (1), community-dwelling individuals with stroke (1), older caucasian participants (2), adult (1) and children (1) tennis players, men with osteopenia or osteoporosis (1), female weightlifters (1) and US female (1) and British male (1) army recruits; for ligament, people studied were female soccer players (1) and male weightlifters (1); for tendon they were basic infantry recruits (1), old female adults (1), participants not strength trained (1), male adults (1) recreationally active volunteers (1) and healthy subjects (4); for muscle, they were premenopausal women with fibromyalgia (FM)(1), young men (1), sedentary men (1), young active men (1) and healthy subjects (3).

Characteristics of intervention

For IVD there was no intervention (observational studies) (4), core exercises (1) and resistance and aerobic exercise (1); for cartilage there was a home exercise program (1) a running program and a marathon (1), aerobic and step aerobic exercise (1), aquatic resistance exercises (1) and a comparison between three types of exercise (2); for bone there was aerobic exercise (1), aqua fitness exercises (1), ordinary curriculum physical activity classes (3), basic military training (2), tennis (1), functional electric rowing (FES) (1), single leg drop exercises (1), high impact unilateral exercises (1), general exercise plus supplementation (1), exercise and hormone therapy (1), multicomponent exercise training (1), odd impact and resistance exercise (1), a combination of resistance, impact and balance exercises (1), high intensity progressive resistance and impact exercises (1), two exercises regimes (1), high-intensity progressive resistance training (1), progressive, moderate-intensity resistance plus impact exercises (2), impact and plyometric exercises (1), no intervention (observational studies)(7) and the rest were a combination of general exercise (7); for ligament both studies were observational, so no intervention as used; for tendon there was basic infantry training (1), plantar flexion exercises (1), isometric plantar flexion (2) explosive isometric unilateral plantar flexion (1), and no intervention (observational studies)(4); for muscle there was total body resistance exercises (1), unilateral resistance exercises (1), leg press exercises (2), bilateral flywheel squats (1), knee extensor exercises (1) and eccentric knee exercise (1).

Comparison

For IVD the comparison was with no sport controls (2), different activity levels (2), core exercises (1) and manual therapy and motor control (1); for cartilage there was a simple control group (1) supplementation (glucosamine sulphate)(1), usual care (1), no training (1) and three types of exercise (2); for bone there was aqua fitness training with control (1) weekly physical activity done in different frequencies (3), between playing arms tennis players (1), tennis in different populations (1), plyometrics in different frequencies (1), drop landing exercises (1), female weightlifters and age matched controls (1), lifetime bone loading (1), exercise plus supplementation (1), between gymnasts and non-gymnasts (1), between types of high intensity training (1), between resistance, balance or jumping exercises (1), FES rowing plus zoledronic acid (1), between exercise and hormone replacement therapy (HRT)(1), a comparison between fine and gross motor activities plus supplementation (calcium)(1), aerobic, resistance and balance exercise compared to upper extremity exercise as control (1), sports with different impact patterns (1), standing hip adduction and abduction, squat and deadlift and a control group (1), resistance exercise, soccer and a control group (1), progressive, moderate-intensity resistance and flexibility training as control (2), lower body resistance plus jumping or upper resistance plus jumping exercise and a control group (1), only exercise training and a control group (1), no comparison (6) and a control group alone (4); for ligament there was no comparison (1) and age-matched controls (1); for tendon there was no exercise as control (2), two exercise regimes with no exercise as control (1), four types of jumping sports (1), elite international level jumping track athletes and healthy recreationally active controls (1), two types of sports and sedentary subjects (1), two types of exercise regime and no specific training as control (1), male runners, female runners and female non-runners(1) and no comparison (1); for muscle there was between men and woman (1), young and old participants (1) concentric and eccentric contractions (2), heavy and low load (1), different velocities and no exercise as controls (1) and healthy women as controls (1).

Outcomes

The most Reported outcomes for IVD were higher IVD in 5 studies and higher T2 values in 3 studies; for cartilage, an increase in thickness (2) and decrease in thickness (2); for bone there were increases in bone mineral density (BMD) (20), increases in bone mineral content (BMC) (10) and increases in cortical area (6); for ligament were increase in cross sectional area (CSA) (1) and increase in volume (1); for tendon there were increases in CSA (6), increases in stiffness (5); for muscle there were increases in mass (4) and increases in CSA (2).

Risk of bias of observational studies per tissue

Characteristics of risk of bias for observational studies are presented in Table 1.

For IVD all criteria were met; only one study had a sample size calculation (Bowden et al., 2018).

For bone all criteria were met; only two studies had non-responders and both had measures undertaken to address non-responders described, response rate described and information about non-responders (Hughes, Gaffney-Stomberg, et al., 2018; O'Leary et al., 2019).

For ligament all criteria were met; only one did not discuss limitations of the study and did not attain ethical approval nor informed consent (Grzelak et al., 2012).

For tendon all criteria were met; three had measures undertaken to address non-responders, response rate described and information about non-responders (Karamanidis & Epro, 2020; Milgrom et al., 2014; Westh et al., 2007) and one had justified its sample size and no clear statistical significance (Milgrom et al., 2014).

All the included studies regarding muscle and cartilage corresponded to experimental designs; therefore, no assessment for observational studies was performed.

Risk of bias of experimental studies per tissue

Characteristics of risk of bias for experimental studies is presented in Table 2.

For IVD one study had a 6/10 score (Khanzadeh et al., 2020) and one had a score of 7/10 (Owen et al., 2020).

For cartilage one study had a 9/10 score (Armagan et al., 2015), one had a score of 8/10 (Munukka et al., 2016), three had a 6/10 score (Benli Küçük, 2017; Cotofana et al., 2010; Koli et al., 2015) and one had a score of 3/10 (Hinterwimmer et al., 2014).

For bone one study had a score of 4/10 (Hasselstrøm et al., 2008) ten records had a score of 5/10 (Detter et al., 2013; Gabr et al., 2016; Greene et al., 2009; Marques et al., 2011, 2013; Milliken et al., 2003; Specker & Binkley, 2003; Vainionpää et al., 2007; Valdimarsson et al., 2006; Wochna et al., 2019), seven had a score of 6/10 (Bailey et al., 2010; Du et al., 2021; Kukuljan et al., 2011; Lambert et al., 2020; Pang et al., 2006; Watson et al., 2015; Winters-Stone & Snow, 2006) and seven had a score of 7/10 (Bolton et al., 2012; Harding et al., 2020; Karinkanta et al., 2007; Lang et al., 2014; Morse et al., 2019; Winters-Stone et al., 2013, 2014).

For tendon three had a score of 5/10 (Arampatzis et al., 2007; Epro et al., 2017; Werkhausen et al., 2018) and one had a score of 6/10 (Bohm et al., 2014).

For muscle four had a score of 5/10 (Bickel et al., 2011; Fernandez-Gonzalo et al., 2014; Franchi et al., 2014; Holm et al., 2008), and three had a score of 6/10 (Franchi et al., 2015; Häkkinen et al., 2002; Marzilger et al., 2020).

The parameters that most of the studies failed to fulfill were allocation concealment, blinding of subjects, blinding of therapists and blinding of assessors. Included studies for ligament tissue corresponded to observational design; therefore, no assessment for experimental studies was performed.

Discussion

This systematic review aimed to summarize the current body of knowledge regarding morphological adaptations in the tissues of the musculoskeletal system in response to chronic exercise and, as a secondary endpoint, to identify the types of exercise best suited to elicit morphological adaptations. A visual summary of the results is presented in Figure 2.

Four studies showed changes in IVD parameters including increased height, higher IVD height relative to vertebral body and better IVD hydration. These adaptations occurred mainly as a response to aerobic exercise. The IVD is divided into two substructures: the nucleus pulposus (NP), providing resistance to compressive forces, and the annulus fibrosus (AF), providing resistance to tension forces. Compression and tension forces are known to trigger anabolic responses such as rearrangement of the cytoskeleton and gene expression (Fearing et al., 2018; Iatridis et al., 2006). Accordingly, in the studies that showed differences in IVD parameters, the population engaged in activities that involved dynamic compression moments in the spine such as jumping, walking or running; it could be the case that aerobic exercise, due to its dynamic compression element, is best suited to elicit a response in the IVD. Two studies did not find changes in IVD height after the intervention. Khanzadeh et al. (2020) didn't find any changes after the intervention consisting of suspension exercises. Although the AF is sensitive to tension forces, it appears that, due to the higher water content in the NP, the responses to exercise are most noticeable in the NP and not the AF. Taking this into consideration, the intervention used might not be appropriate to elicit a response in the IVD. Owen et al. (2020) found differences in IVD height and T2 times, but these didn't reach statistical significance. The authors hypothesize that this may be due to poor exercise adherence and duration of the intervention.

Four studies showed changes in cartilage thickness, volume and a better collagen orientation. These changes were observed as a result of resistance and aerobic exercise. Although not fully understood, it is known that the mechanisms by which the cartilage adapts to mechanical forces is an interaction between the pericellular matrix (PCM) and the chondrocyte. The PCM provides a connection between the extracellular matrix (ECM) and the chondrocyte and houses growth factors and regulatory molecules that bind to the receptors in the chondrocyte in response to mechanical loading such as fibroblast growth factor 2 (FGF2). Meanwhile the chondrocyte is the actual mechanoresponsive organ, housing mechanoreceptors sensitive to the molecules released by the PMC such as ion channels, primary cilia and integrins. This interaction between PCM and the chondrocyte provokes a rearrangement of the cytoskeleton, activation of mitogen-activated protein kinases (MAPK), secretion of Wnts family glycoproteins and expression of MicroRNAs (miRNAs) (Gilbert & Blain, 2018; Vincent & Wann, 2019; Zhao et al., 2020). One study found no changes in cartilage morphology after the intervention. Cotofana et al. (2010) found no change in knee cartilage morphology after three months of intervention. The authors addressed the duration of the intervention as a limitation in their study, which could be the reason for the lack of changes observed, suggesting that three months is too short of a time frame for the cartilage to adapt. In line with this, studies included in this systematic review that observed changes in the cartilage had intervention times of four, six and twelve months. One study found a decrease in cartilage thickness and volume after the intervention period. Hinterwimmer et al. (2014) had a six-month intervention period of marathon supervised training and a marathon run after the intervention period and found a decrease in cartilage thickness and volume. MRI examination times may explain the results observed. The follow up measure was performed one day after the marathon which could represent an acute response to exercise. As an acute response to exercise, cartilage thickness decreases in a dose-dependent manner and the time needed for it to return to baseline levels is unknown (Bini & Bini, 2020; Crowder et al., 2021; Esculier et al., 2019; Harkey et al., 2018; Kersting et al., 2005) so it is possible that the results observed were due to the marathon run. Even considering this the authors interpreted the results as a non-significant decrease in cartilage thickness and volume, indicating that exercise in the form of marathon training and a marathon run is not harmful to cartilage health.

Thirty-six studies reported changes in bone parameters such as increased BMD, CSA, trabecular number, higher cortical area and BMC, with the most reported sites being the hip and spine. These adaptations were observed across different populations and age groups such as healthy people, children, premenopausal women, people with stroke, men with prostate cancer and even individuals with SCI and as a result of resistance, aerobic and plyometric exercises. Bone tissue is known to be sensitive to mechanical loading, triggering different responses such as increasing bone mineral density and strength by stimulation of mesenchymal stem cells, osteoprogenitors, osteoblasts, and the terminally differentiated osteocyte (Yuan et al., 2016). Mechanical stimulation of the bone starts with muscle activity that provokes a slight bend in the bone. This bend initiates fluid flow within the lacuno-canicular network inside the bones which in turn stimulates the osteocyte, triggering bone remodeling (Gusmão & Belangero, 2015; Rosa et al., 2015). Three studies reported no interaction between exercise and bone parameters. Wochna et al. (2019) used a 6-month aqua fitness class intervention in postmenopausal women, and no changes in bone mineral density on the whole body, left hip or spine nor changes in bone turnover markers were observed. The authors reported a small sample size as a limitation in their study which could explain the results obtained. Karinkanta et al. (2007) implemented a 12-month exercise program consisting of resistance training, balance and jumping training, a combination of both, and control group, respectively. Although no changes in BMD were observed, there was an increase in femoral neck strength in both the resistance and combined groups and a 2% less decrease in bone strength index in the combination group compared to the control group at the tibial shaft. The authors hypothesized that this may be due to a redistribution of bone mineral within the femoral neck and tibial shaft in response to exercise and not an increase in bone mineral per se. Greene et al. (2009) used an eight-month drop-landing protocol in prepubertal girls. The protocol consisted of unilateral drop landings with heights of 14 cm and 28 cm. They observed no changes in bone geometry, biomechanical properties nor bone strength index. The authors hypothesized that the reasons no changes were observed were due to the short duration of loading (landing moment), moderate intensity (the relative low heights used) and the lack of variability in movement. The moderate intensity of the drop landings could explain the results observed, as it has been observed in the articles in this review that short duration of loading, like the ones experienced during jogging, and lack of variability in movement, such as the case in plyometric exercises, can provoke changes in bone.

Two studies observed an increase in volume and CSA in the anterior cruciate ligament (ACL) and posterior cruciate ligament (PCL), respectively. It is known that ligament exposure to mechanical loading results in increased stiffness of the tissue (Frank, 2004), although little is known about the specifics of this process. One evidence of mechanical adaptation is the process named ''ligamentization'', referring to a change in histological properties a tendon graft undergoes when used for ACL reconstruction where, overtime, the tendon graft starts to resemble more to a normal ACL than to a tendon. These changes in histological properties have been observed in both animal and human studies and could be explained by the mechanical forces the ligament experiences during normal gait and sports activities (Claes et al., 2011; Pauzenberger et al., 2013). The studies included in this review were observational in design and the tissues studied were the ACL and PCL. It could be that the forces experienced in the knee, which are translated into the ACL and PCL, during squatting or running are what triggers adaptation in the ligament.

Nine studies in this review reported changes in the tendon such as increased stiffness and CSA. These adaptations were observed to occur in the patellar and Achilles tendon, in healthy populations, across different age groups and as a result of resistance and aerobic exercise and different sports such as volleyball, basketball, jumping track, field track, among others. The main mechanosensitive cell in the tendon is the tenocyte. The tenocyte is sensitive to mechanical loads through a series of mechanisms and signaling pathways such as primary cilium, intracellular calcium among others (Lavagnino et al., 2015; Munukka et al., 2016), thus, inducing biochemical changes in tendons that affect the diameter of collagen fibrils and CSA (Wang, 2006). There's still debate whether increased stiffness and CSA may be beneficial or not. For instance, in athletes recovering from patellar tendinopathy, a decrease in stiffness is associated with improved clinical outcome (Breda et al., 2022) whereas other authors have found that greater patellar stiffness may be related to a mechanical improvement in force transmission during muscle contraction (Cristi-Sánchez et al., 2019) and energy storage and release (Wiesinger et al., 2016), and protection from strain injuries (Radovanović et al., 2022).

Seven studies reported changes in the muscle such as an increase in thigh lean mass (TLM), type I and type II fibers, CSA and muscle volume. These adaptations occurred in both women and men, old and young populations and as a result of resistance exercise. Resistance exercise is commonly reported to promote hypertrophy in muscle, with an increase in the muscle CSA (Moghetti et al., 2016). While myosin and actin filaments inside the sarcomere are responsible for muscle contraction, there are a number of possible mechanosensing structures in the sarcomere such as 1) Z-disk, providing a focal point for force transmission between sarcomeres and acting as main anchors during sarcomerogenesis 2) Costameres, acting as focal adhesions inside the skeletal muscle 3) Titin, a protein that lies parallel to the actin and myosin and 4) Filamin-C and Bag3, proteins that reside inside the Z-disk. It has also been proposed that the deformation experienced by the myonuclei as a result of loading of the cytoskeleton, together with the previously mentioned proteins, activates a signaling cascade that triggers adaptation of the muscle (Jani & Schöck, 2009; Wackerhage et al., 2019).

Three studies reported specific adaptations in response to contraction type, in which concentric contractions provoked an increase in pennation angle (PA) while eccentric contractions induced an increase in fascicle length (FL). Although the reasons are not yet understood, and while both types of contractions promote an increase in muscle mass, there seems to be a specific adaptation to contraction type. For instance, as a result of concentric contraction there is an increase of sarcomeres in parallel, thus increasing muscle mid-belly and increasing PA while as a result of eccentric contractions there is an increase in sarcomeres in series, increasing FL (Franchi et al., 2017).

Strengths and limitations

The main strength of this review is that it is the first to summarize the morphological changes in the musculoskeletal system in response to chronic exercise.

Limitations need to be discussed. Firstly, due to the exercise terms often used interchangeably in the literature such as ''endurance exercise'' or ''cardiometabolic exercise'' when referring to aerobic exercise, ''weight training'' or ''strength training'' when referring to resistance exercise and ''physical activity'' or ''physiological loading'' when referring to exercise, it is possible that some articles may not have been included in this review. Secondly, due to the heterogeneity of the studies, such as population or tissue studied and type of exercise intervention, a meta-analysis was not possible.

Conclusions

Exercise has the capacity to trigger morphological adaptations to tissues of the musculoskeletal system. These adaptations are tissue specific and can range from increased thickness, volume, and stiffness and occur mainly due to aerobic, resistance and plyometric exercise. Although it is clear that tissues can adapt, several questions still linger. The minimal dose required to elicit said adaptations is unknown or what type of exercise is best to trigger said adaptations. Although adaptations have been shown to occur in the presence of pathological changes, such as osteopenia and osteoporosis, it is unknown to what degree these tissues can adapt or even if adaptations can occur while the tissue is injured. Furthermore, there's still the question whether these adaptations are harmful, such as a thickened ligament or tendon, or if they can be beneficial, like improvements in functional outcomes in sports or general wellbeing or if these adaptations are constant across lifespan. To advance the practice of exercise medicine, future studies should address these questions.

Funding

No funding was received for this review.

Conflict of interest

The authors declare no conflicts of interest

Los autores de este manuscrito son Licenciados en Terapia Física.

References

Publication Dates

History